1. Field of the Invention
The present invention relates to an image processing apparatus, an image processing method, and a program.
2. Description of the Related Art
The wide-area surveillance camera that takes an image of outdoor etc. is mainly requested for the following three conditions. A first condition to be requested is to image clearly a subject to be noted even in a shooting condition in which the subject is hard to be seen due to the fog, the mist, or the like. A second condition is to image the subject even in a dark environment with low illuminance. A third condition is to obtain a color image because identification by color information, such as color of clothes, may be important in many cases.
Explanations will be given on states satisfying or not-satisfying the above three conditions in a case using a camera with a single-chip color image pickup device of the past.
First of all, the first condition will be explained. It is important that lights reflected on a subject can be reached to an image pickup device of a camera without being decreased in order to image the subject clearly. In a space where the lights reflected on the subject travel to the camera, there existed air (N2 molecule, or the like), moisture such as a fog, airborne particles such as smog, sand or the like. These objects scatter the lights and prevent a clear shooting opportunity.
According to Rayleigh scattering formula where the equality is attained if the airborne particle is small as an air molecule of 0.37 nm in diameter or a portion of smog, a scattering intensity is proportional to the inverse 4th power of the light wavelength. Moreover, the Mie theory will be applied to a grain size of fog, or the like, which is about 1000 nm to 50 μm in diameter, and the scattering intensity is proportional to about in the range of the 0 to the inverse square of the light wavelength. In the visible light of wavelength 400 nm to 700 nm, wavelengths of blue or green light in particular tend to be more scattered, and wavelengths of blue or green light don't reach a camera. Therefore, an image took by a camera that receives visible lights is close to a state visible to the human eye, and the subject in the distance will be hazy with low visibility.
Next, the second condition will be explained. Since the single-chip color image pickup device is mounted with a red, green, blue color filter in front of the photodetector, the sensibility of the photodetector of each color of the red, green, and blue will be lower than the sensibility of a monochrome image pickup device. There is a method (for example, a night-mode imaging) for temporarily saving an IR cut-off filter set in front of an image pickup device and for illuminating the subject by infrared light in order to compensate low sensibility in imaging in a dark environment with low illuminance. However, in an infrared imaging, it is difficult to image a color image, and it is also difficult to improve the sensitivity substantially since it has to be passed through the red, green and blue filter in front of the photodetector in the infrared imaging.
On the other hand, a single-chip monochrome image pickup device is capable of imaging an image with higher sensitivity and higher spatial resolution than the single-chip color image pickup device. However, since the monochrome image pickup device is not capable of obtaining color information, the above-mentioned third condition is not to be fulfilled.